This is from the other post I just made on WB stuff and the info if from TechEdge itself.
As noted above, the basic WB unit's Vout may vary slightly from the specifications shown here. The stoichiometric Vout of 2.50 may be a little higher or lower (+/- 0.03 Volts) and the free-air Vout may be a little higher or lower than 4.00. It's worth noting that the most important calibration value is to get the difference between the stoic Vout (2.50) and free-air (4.00) to be as close as possible to 1.50 Volts so that the interesting tuning AFR range (AFRs of between 10 and 15 for petrol) are then most accurately represented. When this 1.50 volt specification is achieved, and the unit's offset from the nominal 2.50 volt stoic value is compensated for, quite reliable AFR values can be obtained from the unit.
There are many variables. The WB components on the board, the sensors will all vary ever so slightly, and the TechEdge display itself.
Then you have to consider the age of Dig's sensor compared to yours. He might have been using his for years whereas yours is brand new(goes back to the sensor variable). The sensor's readings will change with age. No way around it.
Also remember that the TechEdge display was a modified unit. According to the documentation, it used a 2.49 volt reference and then used different resistors for low end and high end limits and then balanced them using a 1K resistor. The dots or bars(depending on which you chose) don't even use the actual raw data. They use a lookup table programmed in the PIC.
If you take all the "fudging" into account to make the display work, I find that the tolerances you have shown are very good in fact. Everything considered.
I talked to Dig a while back and he doesn't use the TechEdge so he may not be able to answer your questions on the TechEdge display. Don't mean to talk for you, Dig. Just "putting 2 and 2 together".
None of this was to take anything away from Dig's selfless efforts or Peter at TechEdge. All the info is based on a reference. You change that initial reference ever so slightly and the results will change.
Like I said, with all things considered, I think your end results (comparing the TechEdge modified display and Dig's program modifications) are quite remarkable in that they are that close.
JMO,
Blake